COGNITIVE AND BEHAVIORAL PRACTICE 1, 25-45, 1 9 9 4

The Therapeutic Relationship in Cognitive-Behavioral Therapy: Patient Perceptions and Therapist Responses

Jesse H. Wright University of Louisville School of Medicine Louisville, Kentucky

Denise Davis Vanderbilt University School of Medicine and Vanderbilt School of Arts and Sciences Nashville, Tennessee

A working relationship between the patient and therapist is an essential part of any psychotherapy, yet few guidelines exist for this component of cognitive- behavioral treatment. Findings of therapy process and outcome research suggest that the therapeutic relationship strongly influences treatment results, and that interpersonal factors and technical applications interact in forming an effective alliance. Considering the perspective of the patient, we identify general expecta- tions and individual differences that can have an impact on the therapeutic rela- tionship. Individual differences are discussed in four areas of clinical interest: situ- ational concerns; the effects of Axis I psychiatric disorders; sociocuhural influences; and personality structure and schemas. Efforts to understand the patient's per- spective, based on an analysis of general expectations and individual differences, can assist therapists in optimizing the relationship component of cognitive- behavioral therapy.

Each t ime we see a patient, we face the challenge of bu i ld ing or ma in ta in ing a un ique therapeut ic relat ionship. How do we focus our a t tent ion to make sure that this happens, par t icular ly in a form of therapy explicit ly geared toward the

25 1077-7229/94/025-04551.00/0 Copyright 1994 by Association for Advancement of Behavior Therapy

All rights of reproduction in any form reserved.

26 W R I G H T & DAVIS

patient's behavioral changes? What types of self-instructions and therapist be- haviors are most useful in this endeavor? What facilitates the rapport between patient and therapist that is generally regarded as an essential feature of suc- cessful psychotherapy (Beck, Rush, Shaw, & Emery, 1979; Beitman, Goldfried, & Norcross, 1989; Frank, 1985; Strupp, 1988)? How do we approach conflicts in our attempt to create a helpful interaction? Although there are no absolute answers to these questions, empirical investigations and clinical observations suggest that there are ways to promote the formation of good patient-therapist relationships. In this paper, we consider literature on therapy process and out- come research, draw on our clinical experience to discuss general patient expec- tations, and present a conceptualization of individual differences as a frame- work for approaching the many variations among patients. Our goal is to integrate these observations in a way that will help readers maximize their effectiveness in working within a cognitive-behavioral therapy relationship.

Research on the Treatment Relationship

Has An Ideal Approach Been Identified? Therapy process and outcome research has validated the importance of ac-

ceptance, positive regard, and empathy that Rogers (1957) first recognized sev- eral decades ago. A large base of empirical investigation across theoretical do- mains has given us a more refined understanding of these concepts in psychotherapy. Looking first to the "bottom line," the successful outcome of psy- chotherapy, we find strong consensus in the conclusion that the relationship is central to therapeutic change (Beitman et al., 1989; Lambert & Bergin, 1992; Safran, McMain, Crocker, & Murray, 1990). In an analysis of nine major re- view studies of relationship variables, Patterson (1984) concluded that the evi- dence in favor of the "necessity if not sufficiency" (p. 437) of positive relationship conditions was as strong as any in the field of psychology. The most recent and comprehensive review of the literature (Orlinsky, Grawe, & Parks, 1994) con- sidered over 2,300 empirical studies and concluded that the most important de- terminant of outcome is the quality of patient participation in therapy. Further, the therapeutic relationship is a significant mediating link in the process and outcome interaction. Therapist warmth, empathy, and a positive relationship bond are most strongly associated with treatment outcome when the patient's perceptions of these qualities are considered (Lambert & Bergin, 1994; Orlinsky et al., 1994). Even in specific behavioral therapies, patients who view their ther- apist as warm and empathetic will be more involved in their treatment and ulti- mately, have a better outcome (e.g., Holtzworth-Munroe, Jacobson, DeKlyen, & Whisman, 1989; Miller, Taylor, & West, 1980; Williams & Chambless, 1990). According to Orlinsky et al. (1994), these conclusions are based on literally hundreds of empirical results and can be considered scientific "facts" (p. 361).

Although the therapeutic relationship has been defined and measured in sev-

THERAPEUTIC RELATIONSHIP 27

eral different ways (e.g. Alexander & Luborsky, 1986; Horvath & Greenberg, 1986; Safran & Wallner, 1991; Suh, Strupp, & O'Malley, 1986), in simple terms it refers to the personal qualities of the patient, personal qualities of the ther- apist, and the interactions between them (Beitman et al., 1989). The quality and degree of patient collaboration is stronger than any specific therapist vari- able in predicting successful outcome (Hartley, 1985; Stiles, Shapiro, & Elliot, 1986). The patient's role in the treatment relationship appears to be one of a highly influential contributor rather than a passive recipient. Therapists cannot directly alter patient variables, but they have control over how they approach or engage the patient. Therapists contribute to a successful relationship by en- gaging the patient with empathic, affirming, and collaborative interactions that incorporate the skillful application of potent interventions (Orlinsky et al., 1994). Because psychotherapy is a complex change process, however, full appreciation of how it works is unlikely to be accomplished through the discovery of simple associations between variables (Goldfried, Greenberg, & Marmar, 1990).

A conceptualization originally proposed by Bordin (1979) offers a means of integrating general relationship factors, therapeutic techniques, and the pro- cess of interaction. This conceptualization proposes three interacting compo- nents in the therapy relationship: the emotional bond between the patient and therapist; their agreement on the tasks of treatment; and their shared view of the desired outcome or goals of treatment. The bond affects the degree of agree- ment and this in turn affects the quality of the emotional bond. Thus, technical and relationship factors are interdependent parts within a single process in psy- chotherapy (Safran et al., 1990). In a similar vein, other investigators (Butler & Strupp, 1986; Henry, Schacht, & Strupp, 1986; Safran, 1990) have argued that interpersonal interaction is the context in which techniques gain meaning and effectiveness. Behaviorists have traditionally emphasized the use of specific technology rather than interpersonal sensitivity to influence patients. Yet we can clearly see that there are times when the patient's desire to please the ther- apist is the only motivation to initiate or continue behavioral interventions. When the patient likes the therapist and trusts his or her judgment, the patient is more apt to fully engage and persist in the tasks of therapy. On the other hand, we recognize that we can't fully understand the therapy relationship by viewing it in isolation from the process of accomplishing therapeutic tasks and goals. The cognitive-behavioral therapist offers access to an effective, empirically based technology which is then applied in the context of a reinforcing human encounter.

Researchers in the psychodynamic tradition have pinpointed some valuable information about the interpersonal processes of therapy. Process research with short-term psychodynamic psychotherapy by Strupp and colleagues (Suh, O'MaUey, Strupp, & Johnson, 1989) has found that a pattern of interaction es- tablished within the first three sessions of contact is predictive of treatment out- come. Strupp's research indicates that if the therapist can respond to negative patient behaviors with an increase in warmth within the first few sessions, out-

28 W R I G H T & DAVIS

come can be improved for patients who show a low level of interpersonal func- tioning. However, negative or hostile behavior from the therapist tends to pro- duce a downward course from which it is difficult to recover.

The process of interpersonal interactions in dynamic therapy has been ana- lyzed using a fine-grained procedure of coding discrete units of patient-therapist exchanges (Structural Analysis of Social Behavior [SASB]; Benjamin, 1974; 1982). This procedure is based on a circumplex model with two dimensions, control and affiliation, that are thought to characterize interpersonal transactions (Kiesler, 1983). Control refers to the degree to which the conversation is actively directed, and affiliation refers to the degree of warm emotional expression. Using the SASB form of analysis, the degree of fit between patient and therapist is termed complementarity. A complementary fit occurs when there is reciprocity on control (dominance from one is followed by submission from the other and vice versa), and correspondence on affiliation (friendliness followed by friendli- ness, and hostility followed by hostility). Negative complementarity is a hostile and controlling exchange, while positive complementarity is a friendly, autonomy enhancing exchange (Henry et al., 1986). Positive complementarity has been associated with a higher degree of desirable patient changes than negative com- plementarity, or mixed communications (Henry et al., 1986; Svartberg & Stiles, 1992). In the case of mixed communication, these authors suggest that ther- apists should attempt to respond in complementary fashion only to the positive portion of the patient's message.

Early results from a series of investigations of the interpersonal process in dynamic and cognitive modalities suggest that control has a stronger relation- ship to outcome than affiliation in dynamic psychotherapy (Muran, Samstag, Jilton, Batchelder, & Winston, 1992). In an overview of this research program, Muran (1993) noted that friendliness between patient and therapist was posi- tively related to the quality of the alliance, but hostility was negatively related to both the alliance and outcome. An analysis of changes within sessions re- vealed that reciprocal shifts toward a moderate level of therapist control and patient compliance were characteristic of good outcome cases. Thus, it was con- cluded that the most productive interpersonal process in psychotherapy can be characterized as "a didactic interaction against a stable backdrop of friendli- ness" (p. 70). A replication study using cognitive-behavioral treatment is appar- ently underway.

Problem resolution within the therapy process has been the focus of other investigations in this research program. Based on a conceptualization of a dys- functional cognitive-interpersonal cycle (Safran, 1990; Safran & Segal, 1990), problems or ruptures in the therapy alliance are viewed as a unique opportu- nity for assessing the patient's pathogenic beliefs. A model for corrective cogni- tive intervention utilizing the context of the therapy relationship (Safran et al., 1990; Safran & Segal, 1990) is currently being studied. This is an exciting re-

THERAPEUTIC RELATIONSHIP 29

search forefront that may further our understanding of the relationship in cognitive-behavioral therapy.

A recent case history report provides a valuable example of the interaction between a patient's beliefs, the treatment alliance, and therapy process. The pa- tient, a young woman referred for psychological treatment of dysphagia, remained convinced that she had a medical problem even though extensive evaluation revealed no physical abnormalities (Connor-Greene, 1993). Despite concerted efforts by her therapist to provide warmth and encouragement, as well as a spe- cific behavioral treatment plan and clear treatment rationale, no effective alli- ance was established. This was not due to interpersonal transaction problems with the therapist, but rather the mismatch between their respective conceptu- alizations of the problem (medical versus psychological). When the therapist considered an alternate view of the origin of the patient's problems, a dramatic shift occurred in the therapy alliance. The new conceptualization, involving a repressed t rauma from a childhood experience with the Catholic religious ritual of throat blessing on the feastday of St. Blaise, provided the patient with an ex- planation that was more acceptable because it was within the network of beliefs reinforced by her social network and experience. Once this formulation was ac- cepted by the patient, the original treatment plan of exposure was rapidly im- plemented, and there was a dramatic improvement in symptoms. An interesting aspect of this case is that it was the therapist's detailed knowledge of a certain sociocultural experience, Catholic rituals, that yielded an opportunity to produc- tively engage the patient. As Connor-Greene notes, this improvement did not fit the usual clinical picture of recovery from a past t rauma because the patient never actually remembered any trauma. The relevance of the formulation was in reframing the problem in a way that prompted the patient to work directly on her symptoms.

Our discussion of research on the treatment alliance highlights the impor- tance and complexity of the therapy relationship. There are several findings rele- vant to clinical practice. First, the therapy relationship and therapeutic tech- nique are not separate domains, but rather integrated aspects of a single process. Second, the quality of the patient's involvement is crucial to the outcome of psy- chotherapy. Third, there is convincing evidence that therapist warmth and friend- liness play a substantial role in forming an effective therapeutic bond, that in turn mediates patient involvement. Fourth, the degree of positive complemen- tarity between patient and therapist in areas such as control and affiliation is important. Poorer outcome has been associated with hostile, overcontrolling interactions. Although there is some recent evidence that moderate levels of ther- apist control and patient compliance may be associated with improved treat- ment outcome, these findings await replication. In addition, problems or difficul- ties in the alliance can be viewed as prime opportunities for assessing and intervening with maladaptive beliefs and behaviors (Safran et al., 1990; Safran

30 W R I G H T & DAVIS

& Segal, 1990). Finally, Bordin's (1979) formulations, originally developed for analytically oriented therapy, can be adapted for conceptualizing the interac- tion between relationship factors and treatment techniques in cognitive-behav- ioral therapy.

We agree with Bordin and others (Butler & Strupp, 1986; Connor-Greene, 1993; Goldfried et al., 1990; Henry et al., 1986; Safran, 1990) that the therapeutic alliance is a complex phenomenon that influences and is influenced by treat- ment tasks, goals, and content. Thus, our efforts to fully engage patients in ef- fective treatment relationships can be facilitated by recognizing and responding to the broad context of factors that can influence the therapeutic bond. The re- maining sections of this paper discuss methods of improving the treatment alli- ance by increasing our sensitivity to patient perceptions, and responding to situ- ational, diagnostic, sociocultural differences, and individual core belief patterns.

General Expectations for Treatment

What Do Patients Want in a Therapeutic Relationship? One very useful, simple method for assessing patients' general expectations

is an open-ended inquiry. In developing training materials for therapists, we have asked patients to discuss, occasionally on videotape, what they would most like mental-health practitioners to learn about them. These inquiries have been drawn from more than two dozen outpatients in a university medical center mood disorder clinic, a population that is racially diverse but mostly white, middle class, and approximately 40% male and 60% female, who have sought help for disorders of anxiety and depression. This group of patients appears to be similar in several respects (gender, age, race) to the general population who seek mental health services (Vessey & Howard, 1993). The following is a composite list from those candid answers, presented as a hypothetical letter from a patient who outlines common expectations of individuals seeking outpatient treatment. In our experience, other groups of patients (e.g., more severely disturbed inpa- tients) have similar expectations; however, they may have less ability to articu- late their desires. In this letter, the patient informs us about some basic qualita- tive requirements from his or her perspective.

Dear Dr. Wright and Dr. Davis, In response to your request, I am writing to share my

thoughts with you about what therapists need to learn from the patient's point of view. I've made a list of 10 things that a therapist can do that will make a big difference in the satis- faction of their patients. Maybe each one isn't equally im- portant to every person, but I think that most people I know would want their therapist to provide these things.

First, provide a safe and professional setting for our

THERAPEUTIC RELATIONSHIP 31

meeting. The place we meet must be physically safe, pri- vate, confidential , free from distractions, and comfortable. Your office envi ronment tells me a lot about you. I t should communica te a sense of friendliness and authority, en- couraging me to feel welcome and willing to accept what you have to say. Occasional problems or distract ions will usual ly be tolerated (e.g.. cooling system breaks down, un- expected fire drill sounds), but too much of this will even- tual ly bo ther me. A sloppy envi ronment may say that ei- ther you don't really care about me, or that you are not very careful or up to date in your methods. I am especially con- cerned about how you handle informat ion about me, and I want to be sure that you will protect my confidences. So be careful about how you handle files, phone calls, and hallway conversations.

Second, treat me with respect as a person. Show regard for my digni ty a n d my sensitivities. Don' t assume that it is acceptable for you to touch me or behave in an overly personal m a n n e r toward me. Us ing terms such as "hon" or "sweetheart" or other nicknames in reference to me is overly personal. But, if you're too dis tant from me, it will seem that I 'm t reated like an object. I f you avoid normal things like saying hello and goodbye, I'll th ink that you are dis tant and unfr iendly to me. Paying too much at tent ion to my looks, my weight, my status and title (or my lack of status and title) can all lead me to think that you don't really re- late to me as a whole person.

Third, take my concerns seriously. Don't minimize them, ignore my pain, or brush me off as a hysterical patient. I f I b r ing a p rob lem or idea up for discussion, don't dismiss it as un impor tan t . I t is d iscouraging to get the impression that you think I am being silly or that I want to feel bad. I f you brush me off, ! may think that you don't want to take the t ime or effort to unders tand my prob lem and that you

may not provide adequate t rea tment . Fourth, I want to think you have my best interests in

mind. W h e n it seems like your research or busy schedule comes first, my feelings might be hurt . I don't want to be thought of as a n u m b e r or a case to be crossed off a list. I f I've te lephoned you with a quest ion or request, I need to believe that you consider me impor tan t enough to re- turn the call, even if it is not an emergency.

3 2 WRIGHT & DAVIS

Fifth, I want you to know what you're doing. I don't neces- sarily require you to be the world's top expert, but I want to think that you are good at what you do and that I can trust your judgment . I would like you to provide me with informat ion about your ideas for t rea tment and to explain the basis for your thinking. At the same time, I want you to recognize your l imits and let me know when you are not informed about something. I appreciate your efforts to learn new things and hope that you will share anything that could be useful to me.

Sixth, give me practical information. Let me benefit from your experience in working with other people who have had similar problems. Don't hold back ideas about solutions that might help me make improvements in my life. I might not discover these solutions on my own. If you make sugges- tions, help me unders tand how to carry them out and how I might solve obstacles that could get in the way.

Seventh, allow me to make choices with your informa- t ion and suggestions. Don' t always push your ideas on me. Help me learn about the benefits and risks of different choices, but don't assume you know what is best for me. I want to feel s tronger in myself and not jus t reflect confi- dence in your j u d g m e n t or your abi l i ty to persuade me. I am the one who will live with the outcome of my choices.

Eighth, stay flexible in your th inking about me. Don't allow your theory to determine everything about me, regard- less of how well it may fit. You may be certain you have a clear under s t and ing of my problems, but it's possible that you could miss something that 's very impor tant . Give some weight to my circumstances and how other people act apar t from me. Don' t automat ica l ly locate all of the problems within me. Sometimes my symptoms are the only way I have found to cope with a difficult situation.

Ninth, follow up on your recommendat ions . Ask about whether I have followed your suggestions and check to see what happened. This tells me you care about what I do. I f something isn't p roduc ing the desired results, I want you to make another suggestion. Be sensitive to any possible dis- comfort, even if I say that I'll go along with what you ask me to do. I f something is uncomfor table or distasteful to me, I want you to consider other options and not neces- sarily th ink I am a bad or uncooperat ive patient.

Tenth, pace yourself. I f you are overworked, unhappy,

THERAPEUTIC RELATIONSHIP 33

or tense, I may think you aren't being a very good example. I may resent your rushing, being distracted, or having fre- quent schedule changes. I may feel inferior because I see myself as less successful or important. I may try to reverse roles and take care of you. I f this happens, I'll fail to get what I need from treatment. I may think you're not doing your job with me as well as you could. More than anything, I may get discouraged and think your techniques must not work very well if even you can't successfully use them.

Best wishes in your efforts to help other therapists learn about how patients want to be treated.

Sincerely, Your patient

None of these expectations are specific to cognitive-behavioral therapy, but all involve possible variations in therapist behavior. They reflect an attention to what is often called the "non-specific" aspect of therapy. As such, they can affect each of the components of the therapy alliance proposed by Bordin (1979): the emotional bond, and the degree of agreement on goals and tasks. We sug- gest that the sooner and more completely these general expectations can be met, the more smoothly therapy will progress.

Individual Differences

How Should Therapists Respond to Patient Variability? Along with considering the patient's general expectations for therapy, ther-

apists face the challenge of responding to individual differences in expectations and perceptions of the therapy interaction. Expectations for therapist behavior can vary widely. For example, some of our patients may want us to be authorita- tive, directive, and active; and others would prefer a nondirective and relatively inactive therapist style. The latter characteristics, nondirective and relatively inactive, seem to fit poorly with typical cognitive-behavioral interventions. Sev- eral possible dimensions of patient expectations for therapist behavior are listed in Table 1. These dimensions are derived from clinical experience and are presented simply as a reminder of the wealth of diversity in patient attitudes toward the therapeutic relationship. Many other expectations or requirements for therapist responses are possible.

Individual differences in the patient's view of the treatment relationship may be due to a number of potentially interlocking factors such as: temperament, the type and severity of clinical symptoms (e.g. hostility and irritability in a bipolar patient or ruminations in an obsessive/compulsive), level of intelligence, sociocultural influences, experiences of friends or family in psychiatric treat- ment, social skills capabilities, and underlying schemas. With such a wide va-

34 W R I G H T & DAVIS

TABLE 1 INDIVIDUAL DIFFERENCES IN THE THERAPEUTIC RELATIONSHIP: DIMENSIONS

OF PATIENT EXPECTATIONS FOR THERAPIST BEHAVIOR

Authoritative-Unassuming Directive-Nondirective Structured-Unstructured Friend-Professional Active-Inactive Humorous-Reserved Upbeat/Hopeful-Cautious/Reserved Same Sex-Opposite Sex Didactic-Nondidactic Expressive-Controlled

riety of influences, how should the cognitive-behavioral therapist attempt to un- derstand and manage variability in patients' reactions to the therapeutic relationship? We suggest that, in addition to responding to general expectations, at least four areas of clinical interest that are discussed in detail below be consid- ered: (1) situational issues; (2) impact of psychiatric disorders (Axis I diagnoses); (3) sociocultural influences; and (4) personality structure and individual belief patterns. We suggest that examining these four areas of individual differences may reduce the risk for fundamental attribution errors (Nisbett & Ross, 1980) such as mislabeling a response as indicative of enduring personality traits when it is due to an understandable situational concern. Proceeding from the first through the fourth level of this method of examining the treatment alliance would lead to a sequential deepening of the therapist's understanding of relationship variables. However, in clinical practice it may he best for the therapist to simul- taneously consider information from all four areas as it becomes available. In- fluences of individual beliefs can be evaluated in more detail as it becomes avail- able. Influences of individual beliefs can be evaluated in more detail as these underlying cognitive structures emerge in the therapy process.

Situational issues. External, situational factors constitute a potential influence on the therapy alliance that has not been studied or even sufficiently acknowledged (Garfield, 1992). Sometimes, patients' expectations or perceptions are indica- tive of highly salient aspects of their current life situation (Kranz, 1985). Situa- tional concerns or factors may activate a certain response "set" that is not neces- sarily characteristic of the patient at other times. To manage situational influences, it may be best to consider specific patient requests on a practical or "surface" level. This is especially true in the initial phase of treatment when an attempt to understand the patients' perceptions by uncovering belief systems may be premature. A deeper level of assessment and response to idiosyncratic expecta- tions can be delayed until a solid therapeutic relationship is established, and more data are available.

THERAPEUTIC RELATIONSHIP 35

For example, one of our patients made it known in the first session that she was very careful about finances and would expect to receive a reduced bill if for any reason the therapist was late or needed to shorten the session. Another patient requested that his therapist never call him at work or home because he wanted no one to know about the psychotherapy. Both of these examples could represent behaviors indicative of dysfunctional beliefs, but early in treatment, the patient's expectations often may be best managed on a pragmatic basis.

A straightforward, natural response can be used to help build an effective therapeutic relationship. In the case of the patient who requested a fee reduc- tion, the therapist simply agreed that this would be reasonable, but that every attempt would be made to meet for the entire time scheduled. The therapist also used this interchange in the development of the case formulation and hypothesized that later work on beliefs related to trust and control might be useful. It wasn't long, however, before the therapist found out that the patient's concerns stemmed from circumstances involving a newly diagnosed chronic ill- ness in a family member and a huge, unexpected tax bill due to an accounting error in the family business. Issues of trust and control were largely circum- scribed to these problems, and did not persist as a theme in the therapy rela- tionship.

Patients who do not articulate their expectations in the beginning of treat- ment generally present a larger problem than those who do. We have found that this is particularly true with the reluctant patient: one who has been admitted to the hospital abruptly without really considering the need for treatment, or one who is forced into therapy by an angry spouse or by legal decree. Yet, even the willing and seemingly well prepared patient can come to therapy with ex- pectations that are not immediately apparent. Difficulties with not meeting ex- pectations or initiating strategies that run directly counter to the patient's wishes can be minimized if the therapist: (1) listens carefully to the patient; (2) makes it clear that feedback about the therapist's actions is desired; (3) asks for feed- back at every session; (4) devotes time for setting consensual goals for treatment at the beginning of treatment and for selecting agenda topics at each session; (5) is alert to common themes in patient requirements for therapy; and (6) re- mains flexible and nondefensive throughout therapy.

At times patients may have expectations that are clearly unrealistic or poten- tially damaging, no matter what the situational antecedents. These expecta- tions need to be identified, and clear boundaries should be established for the therapeutic relationship. The most obvious and dangerous expectation is for extensive friendship or physical intimacy. When these thoughts are detected, it is important for us to draw firm guidelines that protect both the patient and ourselves (Gutheil & Gabbard, 1993). Other expectations with possible nega- tive consequences include demands for regularly extending the length of ses- sions beyond the scheduled time or having minimal if any limits on telephone access to the therapist. This can be managed with an explanation of the bound-

36 W R I O H T & D A VIS

aries and limits for a psychotherapy relationship, but an assessment of specific underlying beliefs and patterns of behavior may be required if these expecta- tions or behaviors persist. An understanding of influences of specific psychiatric disorders, as described in the following section of the paper, also can help the therapist respond to maladaptive expectations for the therapeutic relationship.

Impact of psychiatric disorders. Axis I conditions such as major depression, bipolar disorder, or eating disorders may be associated with substantial differences in the patient's expectations for therapist behavior. These differences can be par- ticularly acute when the illness is at a high level of intensity. For example, a patient with severe major depression is likely to be hopeless, suicidal, and para- lyzed with profound psychomotor retardation. Severely depressed patients usu- ally are desperate for help and looking for someone to release them from their suffering. Because they've run out of solutions and feel exhausted, they need a therapist who can inject hope and energy into the therapeutic relationship. Thus, we may need to do much of the patient's work in the early part of therapy (Ludgate, Wright, Bowers, and Camp, 1993; Thase & Wright, 1991).

Patients with psychotic disorders present a special challenge to the cognitive therapist. Psychotic patients can be hostile, irritable, guarded, agitated, or with- drawn. Expectations for therapist behavior may be highly variable between different patients and also within individuals across time (Kingdon & Turkington, 1994). However, persons with psychotic illnesses have the capacity to develop stable and productive therapy relationships (Kingdon & Turkington, 1994). Schizophrenic patients who attend a long-term aftercare group of one of the authors seem to especially value consistency and personal warmth. They also want pragmatic help in managing problems like living with aging parents or coping with social isolation and boredom. This group wants the therapist to be fairly active and to use gentle humor in the therapeutic approach. The latter therapist behavior is one of the groups' favorite mechanisms for reducing ten- sion and obtaining a healthy perspective on the rigors of living with a chronic psychosis.

Acutely ill psychotic patients can have very different expectations. If para- noia is significantly active, they may want us to go away. However, other pa- tients with acute psychosis simply want help with the painful symptoms of the disorder. It has been recommended that therapists should begin cognitive therapy with psychotic patients by attempting to establish a therapeutic relationship and then using a "normalizing rationale" to explain the disorder (Kingdon & Tur- kington, 1991). According to these guidelines, the therapist generally should delay certain cognitive therapy interventions such as eliciting and attempting to modify delusions, until a good therapeutic relationship is established and the patient has a basic understanding of the disorder. The use of a collaborative alliance in exploring alternative explanations for delusions, as opposed to a more authoritative stance, has been emphasized in recent articles on cognitive therapy

THERAPEUTIC RELATIONSHIP ~7

for psychotic illnesses (e.g., Bentall, Haddock, & Slade, in press; Chadwick, Lowe, H o r n e , & Higson, in press).

Other examples of how individual differences in psychiatric disorders might affect the therapeutic relationship include the tendency of panic or phobic dis- order patients to want help in avoiding anxiety; the expectation of eating dis- order patients that therapists will confirm their maladaptive attitudes; and the wish of alcohol abusing patients to have a therapist who can be deceived or will go along with rationalizations and denial. Sensitivity to influences of specific psychiatric conditions on the therapeutic relationship and a knowledge of special- ized CBT procedures for major psychiatric disorders (See for example, Beck, Emery, & Greenberg, 1985; Bowers, 1993; Kingdon & Turkington, 1994; Ludgate et al., 1993; Scott, Byers & Turkington, 1993) can help us adjust our actions to maximize the chances of an effective agreement on the goals and tasks of therapy.

It is also important to meet patient expectations for biological intervention as a possible form of treatment. Many patients with significant Axis I disorders want pharmacotherapy and should be provided with this option. We believe that the therapist has a responsibility to prescribe appropriate medications or to ar- range for medication consultation. Guidelines for managing the therapeutic rela- tionship in combined CBT and pharmacotherapy have been described previ- ously (Wright, 1987; Wright & Schrodt, 1989; Thase & Wright, 1991). A well integrated treatment approach has been recommended in which a cognitive- biological theoretical model is explained to the patient, and the therapists work closely together (in the multiple therapist condition) to deliver a full spectrum of treatment interventions.

Sociocultural influences. The relational bond between patient and therapist can be significantly impaired by negative beliefs or stereotypic attitudes. In this sec- tion of the paper, we focus on reactions that can be traced to an underlying be- lief that reflects a learned sociocultural bias. Such maladaptive reactions can be automatically triggered by broad personal variables including, but not limited to, gender, race, ethnic group, age, socioeconomic status, physical disability, re- ligion, or sexual orientation. Although negative sociocultural beliefs are often deeply held, these may not be very obvious to the person holding them. Either participant in the therapy relationship may be vulnerable to having sociocultur- ally determined biases that influence his or her perceptions of the other person. To facilitate an emotional bond, we must not only be alert to the potential for a patient having negative attitudes toward us (our stimulus value), but to the possibilities of us having similar reactions. As therapists, we might like to be- lieve we are free from prejudices; but we know that having been socialized in a particular culture, we cannot be free from conditioned negative emotional re- sponses, or cognitive biases. We also assume, on the basis of cognitive-behavioral theory, that our schemas or cognitive patterns will actively participate in how

38 WRIGHT & DAVIS

we process information and will direct our behavior as we establish our therapy relationships. This means, as Bernstein (1993) has pointed out in the context of working with gay and lesbian clients, "our work to change our prejudices must be ongoing" From the patient's perspective, it is likely that a therapist's lack of effort in understanding and overcoming biases will be seen as offensive and damaging to the therapeutic bond. The following are some suggested ways, drawn from the behavioral literature, to detect and modify negative biases that are based on social or cultural beliefs.

First, pay attention to your cognitive and emotional reactions to the patient, particularly negative ones. Neglected maladaptive therapist reactions can in- vade and fester in therapy transactions. Once you have detected some negative response such as irritation, dread of the appointment hour, thoughts that label the patient in stereotypic, derogatory terms (e.g., Jewish American Princess), or thoughts of disliking the patient, standard cognitive-behavioral techniques can be useful in pinpointing and adjusting beliefs that are a functional problem in the therapist. Overcoming conditioned negative responses and avoidance be- haviors can best be accomplished with a combination of cognitive restructuring and exposure with response prevention (Spencer & Hemmer, 1993).

Sometimes, however, our negative responses might occur below the threshold of our awareness. We may not attend fully to our own reactions because other concerns compete for our attention. During treatment sessions we may work diligently to keep the tasks of therapy moving forward and thus fail to detect our own emotional state. We also may be inclined to avoid thinking about our negative thoughts and feelings toward patients because of an uncomfortable dis- sonance with our underlying view of ourselves as tolerant, empathic, and politi- cally correct therapists.

Audio and videotaped therapy sessions can serve as an excellent resource in detecting possible underlying biases. By watching or listening to a taped ses- sion, you can become more aware of body language, vocal tone, or speech tran- sitions that might be indicative of stereotyped responses. For instance, do you tend to speak in a loud tone, or enunciate slowly with your patient who has ambulatory limits but full hearing capacity? Do you physically cringe and frown slightly when your gay patients bring up their sex life or their relationships with their children? Do your listening skills sound as discerning with your elderly patients as with younger individuals, or do you tend to abruptly redirect the focus without giving meaningful feedback? Is your eye contact diminished and your vocal tone condescending with an economically disadvantaged patient? Simi- larly, are you hyperattentive with the famous or well-connected patient? When these observations are made, there is an opportunity to further explore under- lying feelings and thoughts the therapist is bringing into the relationship inter- action. Biased attitudes can be challenged and specific behaviors can be cor- rected via self-instructions and practice.

A third way to recognize and challenge social and cultural stereotypes is to

THERAPEUTIC RELATIONSHIP 39

attempt to conceptualize problems from an alternative point of view. In a sim- plified fashion, this involves two basic steps: first, gathering information, con- cepts, and propositions that challenge or expand your existing ways of thinking about the stereotyped group; and second, applying those new constructions in specific clinical situations. A helpful illustration of this method was presented in a case study by Brooks (1993). A woman was referred to Brooks after a suc- cession of (male) therapists had unsuccessfully attempted to treat her, each con- cluding that she was a hopelessly bitter woman and a troublesome wife. Brooks conceptualized her difficulties through the lens of a feminist perspective in an effort to find a workable approach to this "problem" patient. Moving beyond the stereotype of the "typical doctor's wife" that overemphasized pathology in the patient, Brooks considered the sexist elements of her marriage and social arrangements that diminished the patient's self-esteem and led others to with- draw reinforcement and behave in punishing ways toward her. With this con- ceptualization, Brooks was able to feel greater empathy, adjust his style of inter- action, recognize issues in his personal emotional reactivity, and attend to a wider range of concerns in selecting treatment goals.

Effective conceptualization of sociocultural issues also requires having suffi- cient information about specific personal experiences and coping skills that are common in a cultural group. For example, Bernstein (1993), offers a useful dis- cussion of gay and lesbian concerns that are relevant to clinical inquiry. We need to know enough about the patient's unique experiences in order to ask the right questions. And we must know enough about the sociocultural background to understand that certain behaviors that are pathological in other populations could be necessary coping skills for a particular patient. Without this knowledge base, we are more vulnerable to stereotypic distortions that lead to errors in clinical judgment.

There are many ways to become knowledgeable about a group or culture. For example, ideas for understanding the gay and lesbian culture suggested by Spencer and Hemmer (1993) include reading lesbian or gay affirmative publi- cations, attending lesbian or gay sponsored activities and events, or going to a lesbian or gay club or coffeehouse. Other techniques include researching rele- vant professional literature, attending specialized workshops, and consulting with colleagues. Discussions with other mental health professionals are some- times the fastest way to test your conceptualizations, but these individuals must be selected with care. The most obvious risk would be in mutually reinforcing the more emotionally based stereotypes, rather than engaging in a productive exchange of information and critical thinking (see Gambrill, 1990; 1993, for a more thorough discussion of critical thinking and clinical judgment).

The fourth and last way we suggest to actively challenge social and cultural stereotypes is monitoring your clinical atmosphere and making efforts to en- sure that it is welcoming to all categories of patients. Failure to eliminate biases that are evident on clinical forms, in the waiting room reading material, and

4 0 WRIGHT & DAVIS

in the attitudes and behaviors of staff indicates a lack of sensitivity to sociocul- tural influences and a tacit reinforcement of prejudice. On the other hand, caring about these issues establishes an effective starting point for the relationship bond.

Personality structure and individual beliefs. Constructs from CBT can be used to understand variability in patient behavior, and cognitive-behavioral formula- tions can be used to guide us in dealing with vexing problems in therapeutic relationships. Even more importantly, vicissitudes in the therapy relationship can provide us with an opportunity to work directly with the patient's most sig- nificant maladaptive schemas: the cognitive structures that may be at the core of their dysfunction. In the maladaptive cognitive-interpersonal cycle described by Safran (1990) and Safran and Segal (1990), the patient's perceptions of ther- apist behavior provide a phenomenological link to dysfunctional interpersonal schemas and associated patterns of behavior. Although Safran utilizes an inter- personal theory of personality organization, the concept of the cognitive- interpersonal cycle appears to be adaptable to other variations on a cognitive theory of personality. Here we describe an example of individual maladaptive beliefs, based on the Beck et al. (1990) model of personality, that were assessed and modified within the therapeutic relationship.

In the Beck et al. (1990) model, personality is viewed as a combination of core beliefs and a characteristic set of basic strategies for interacting with others. Interrelated beliefs and action plans are organized within a generic knowledge structure referred to in the cognitive literature as a schema. Schemas guide and actively participate in the processing of information and action, so that we see recurring patterns of thoughts, feelings and behaviors. For example, a patient with a dependent personality may have beliefs pertaining to weakness, needi- ness, and incompetence. In contrast, others are seen as strong, powerful, and capable of nurturance (Beck et al., 1990; Fleming & Pretzer, 1990; Turkat & Maisto, 1985. The main behavioral strategy for such an individual may be to cultivate relationships with caretakers and to feel anxious about doing some- thing assertive or contradictory to the wishes of their caretakers. Typical beliefs and behavioral strategies have been described for the full range of personality disorders (Beck et al., 1990). We have found these conceptualizations to be helpful in enhancing therapeutic relationships, as illustrated in the following example.

Mr. D was a 45-year-old banker who initiated cognitive-behavioral therapy because of depression that had begun in the aftermath of a divorce. In the early part of treatment, interventions such as the Daily Record of Dysfunctional Thoughts (DRDT, Beck et al., 1979) appeared to be helpful, and symptoms of depression began to improve. However, it soon became apparent that obsessive- compulsive traits were influencing the therapeutic relationship. At first, Mr. D appeared to be the "perfect" patient. He was extremely polite and apprecia- tive. Homework assignments were carried out with considerable attention to detail. Then, Mr. D began to do much more homework than the therapist as- signed. A request to complete a D R D T would lead to the production in 1 week

T H E R A P E U T I C R E L A T I O N S H I P 41

of a voluminous folder of these forms. Mr. D reported that he was spending 2 or more hours every night reviewing the homework and completing cross refer- ences of his responses to reading assignments. He attempted to inform the ther- apist in minute detail regarding his self-help efforts.

Mr. D's behavior in the therapeutic relationship paralleled his obsessive- compulsive style in other relationships. Beck et al. (1990) have noted that such individuals tend to have personal beliefs related to a heightened sense of respon- sibility and accountability. Patients with obsessive-compulsive personalities have main behavioral strategies of applying excessive rules, acting in a perfectionist manner, placing undue emphasis on evaluating self and others, and attempting to exert control. These beliefs and behaviors had played a large role in the break- up of Mr. D's marriage and also were recapitulated in the therapeutic relation- ship. Mr. D worked extremely hard in therapy and expected the therapist do the same. Notebooks full of homework exercises were brought to the therapist with the request that he digest all of this material between sessions. Despite the patient's high level of activity, there was a sterile quality to the therapeutic rela- tionship. Emotion was rarely expressed by Mr. D, and the therapist was discom- forted by a lack of true collaboration in both cognitive and affective spheres.

This case illustrates the formation of what some authors (Beck et al., 1990; Wright, 1988; Wright & Beck, in press) have termed a cognitive transference. Deeply held beliefs and well ingrained behavioral patterns are repeated in the therapeutic relationship and interfere with the effective use of collaborative em- piricism. The patient's expectations and perceptions of the therapist are typi- cally distorted to remain consistent with his or her underlying beliefs. Behavior also remains consistent with dysfunctional beliefs. In this case, Mr. D expected a high level of productivity from the therapist, was quick to perceive that either he or the therapist wasn't doing "enough", and frequently checked on the ther- apist's progress in reviewing his homework or assimilating the details of his self- help efforts. When a cognitive transference develops, therapists can use CBT procedures to identify and modify the maladaptive beliefs and to work on changing dysfunctional behavior. With Axis II disorders, this may require longer- term therapy than is usually the case with major depression and anxiety disorders.

Treatment with Mr. D extended for a 6-month period and also included oc- casional follow-up visits over the next 5 years. After the first few sessions, the therapist began to focus on distortions evident in the therapeutic relationship that could be targets for therapeutic interventions. The main initiative was directed at uncovering maladaptive beliefs related to perfectionism, responsi- bility, and control, and to demonstrate how the patient's basic attitudes were influencing his relationships. The therapist altered his behavior somewhat in order to "loosen up" the therapeutic relationship and to model more flexible and open interpersonal relations. Humorous interchanges, discussion of the ther- apists' personal views and experiences, and stimulation of affective responses through role play and in vivo exercises were used whenever possible. Mr. D's

42 WRIGHT & DAVIS

depression resolved in the first 4 weeks of treatment, but an extended therapy was required for significant modifications in his obsessive-compulsive behavioral style. Work on the therapeutic relationship was an integral part of the treatment process.

Examination of beliefs, related behavioral strategies, and patterns in the ther- apeutic relationship can help us recognize and manage each patient's unique expectations within treatment. However, if a schema level analysis of the ther- apeutic relationship is ignored, therapists may be baffled by common problems such as non-compliance, excessive dependency, or stagnation of the treatment process. Axis II disorders may be particularly difficult in this respect. Also, in- dividuals who have suffered serious neglect or abuse may show regressive and pathological behavior in the therapy relationship (Liotti, 1991).

We believe that training programs for cognitive-behavioral therapists should include intense supervision on relationship issues. This element of cognitive- behavior therapy may require as much attention as learning how to implement specific behavioral techniques. Our ongoing education as therapists also can benefit from efforts to learn more about how personality structure influences our attempts to form collaborative treatment relationships.

Summary

The therapeutic relationship is an essential, interactive component of cog- nitive-behavioral therapy. We maintain that cognitive-behavioral therapists need to be sensitive to both the general and idiosyncratic expectations of their pa- tients, without compromising the necessary limits or boundaries of the rela- tionship. "Non-specific" elements of treatment can be translated into specific therapist responses for building effective treatment relationships. First, we can strive to be fully attuned to patients' general requirements for therapy, as illus- trated in the hypothetical letter contained in this article. Next, we can closely monitor and shape our own behavior, taking into account our patient's life situ- ation, symptoms, and sociocultural experiences so that we respond in ways that are genuine, friendly, and helpful to them. Finally, we can be aware of possible individual differences in expectations among patients and be prepared to apply cognitive-behavioral theories and procedures as we adjust our therapy style. At- tention to these principles can help us build productive therapy relationships.

References

Alexander, L. B., & Luborsky, L. (1986). The Penn helping alliance scales. In L. S. Greenberg & W. M. Pinsof (Eds.), Thepsychotherapeutic process: A research handbook (pp. 325-366). New York: Guilford.

Beck A. T. (1967). Depression: Clinical, experimental, and theoretical aspects. New York: Harper & Row. Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety disorders and phobias: A cognitiveperspeetive.

New York: Basic Books, Inc.

THERAPEUTIC RELATIONSHIP 43

Beck A. T., Freeman, A., Pretzer, J., Davis, D. D., Fleming, B., Ottaviani, R., Beck, J., Simon, K. M., Padesky, C., Meyer, J., & Trexler, L. (1990). Cognitive thwapy of personality disorders. New York: Guilford Press.

Beck A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy ofdapression. New York: Guilford.

Beitman, B., Goldfried, M., Norcross, J. (1989). The movement toward integrating the psychother- apies: An overview. Artwriean Journal of Psychiatry, 146(2), 138-147.

Benjamin, L. S. (1974). Structural analysis of social behavior. Psychological Review, 81, 392-425. Benjamin, L. S. (1982). Use of structural analysis of social behavior (SASB) to guide intervention

in psychotherapy. InJ. C. Anchin & D. J. Kiesler (Eds.), Handbook ofinterpersonalpsychoth~mpy (pp. 190-212). New York: Pergamon Press.

Bentall, R. P., Haddock, G., & Slade, P. (in press). Cognitive-behavior therapy for persistent audi- tory hallucinations: From theory to therapy. Behavior Th~rtapy.

Bemstein, G. (1993). Assessment and goal selection with lesbian and gay clients, the Behavior Ther- apist, 16, 37-40.

Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research and Practice, 16, 252-260.

Bowers, W. (1993). Cognitive therapy for eating disorders. In J. H. Wright, M. E. Thase, A. T. Beck &J. W. Ludgate (Eds.), Cognitive therapy with inpatients (pp. 337-356). New York: Guilford Press.

Brooks, G. (1993). The gender socialized therapist - A case study. The Psyvhothwapy Bulletin, 28, 15-18. Butler, S. F., & Stmpp, H. H. (1986). Specific and nonspecific factors in psychotherapy: A problematic

paradigm for psychotherapy research. Psychothwapy, 23, 30-40. Chadwick, P. D.J., Lowe, C. F., Home, P. J., Higson, P. J. (in press). Modifying delusions: The

role of empirical testing. Behavior Therapy. Connor-Greene, P. (1993). The therapeutic context: Preconditions for change in psychotherapy.

Psychotherapy, 30(3), 375-382. Fleming, B., Pretzer, J. L. (1990). Cognitive-behavioral approaches to personality disorders. Progress

in Behavior Modification, 25, 119-151. Frank, J. (1985). Therapeutic components shared by all psychotherapies. In M. Mahoney & A.

Freeman (Eds.), Cognition and psychotherapy (pp. 49-79). New York: Plenum. Gambrill, E. (1990). Critical thinking in clinical practice. San Francisco: Jossey-Bass. Gambrill, E. (1993). What critical thinking offers to clinicians and clients, the Behavior Therapist,

16, 141-147. Garfield, S. L. (1992). Major issues in psychotherapy research. In D. Freedheim (Ed.), History of

psychotherapy: A century of change (pp. 335-359)i Washington, D.C.: American Psychological As- sociation.

Goldfried, M. R., Greenberg, L. S., & Marmar, C. (1990). Individual psychotherapy: Process and outcome. Annual Review of Psychology, 41, 659-688.

Gutheil, T. G., & Gabbard, G. O. (1993). The concept of boundary in clinical practice: Theoretical and risk-management dimensions. American Journal of Psychiatry, 150, 188-196.

Hartley, D. (1985). Research on the therapeutic alliance in psychotherapy. American Psychiatric Associ- ation Annual Review, 4, 532-549.

Henry, W. P., Schacht, T. E., & Strupp, H. H. (1986). Structural analysis of social behavior: Appli- cation to a study of interpersonal process in differential psyehotherapeutic outcome. Journal of Consulting and Clinical Psychology, 54, 27-31.

Holtzworth-Munroe, A., Jacobson, N., DeKlyen, M., & Whisman, M. (1989). Relationship be- tween behavioral marital therapy outcome and process variables. Journal of Consulting and Clin- ical Psyvhology, 57, 658-662.

Horvath, A., & Greenberg, L. S. (1986). The development of the working alliance inventory. In L. S. Greenberg & W. S. Pinsof(Eds.), Thepsyehotherapeuticproeess: A research handbook (pp. 529-556). New York: Guilford.

Kiesler, D. J. (1983). The 1982 interpersonal circle: A taxonomy for complementarity in human transactions. Psychological Review, 90, 185-214.

Kingdon, D., & Turldngton, D. (1991). The use of cognitive behavior therapy with a normalizing rationale in schizophrenia. Journal of Nervous and Mental Diseases, 179, 207-211.

Kingdon, D., & Turkington, D. (1994). Cognitive-behavioral therapy of schizophrenia. New York: Guilford.

44 WRIGHT & DAVIS

Krantz, S.E. (1985). When depressive cognitions reflect negative realities. Cognitive Therapy and Re- search, 9, 595-610.

Lambert, M. J., & Bergin, A. E. (1992). Achievements and limitations of psychotherapy research. In D. Freedheim (Ed.), History of psychotherapy: A century of change (pp. 360-390). Washington, D.C.: American Psychological Association.

Lambert, M. J., & Bergin, A. E. (1994). The effectiveness of psychotherapy. In A. Bergin & S. Garfield, (Eds.), Handbook of psychotherapy and behavior change (4th ed.), (pp. 143-189). New York: Wiley.

Liotti, G. (1991). Patterns of attachments and the assessment of interpersonal schemata: Under- standing and changing difficult patient-therapist relationships in cognitive psychotherapy. Journal of Cognitive Psychotherapy: An International Quarterly, 5, 105-113.

Ludgate, J. W., Wright, J. H., Bowers, W., & Camp, G. E (1993). Individual cognitive therapy with inpatients. InJ . H. Wright, M. E. Thase, A. T. Beck, &J. Ludgate (Eds.), Cognitive therapy with inpatients: Developing a cognitive mileau (pp. 91-120). New York: Guilford.

Miller, W. R., Taylor, C. A., & West, J. C. (1980). Focused versus broad-spectrum behavior therapy for problem drinkers. Journal of Consulting & Clinical Psychology, 48, 590-601.

Muran, J. C. (1993). The self in cognitive-behavioral research: An interpersonal perspective, the Behavior Therapist, 16, 69-73.

Muran, J. C., Samstag, L. W., Jilton, R., Batchelder, S., & Winston, A. (1992). Relations of interper- sonal behavior and transactions to alliance and outcome over time in short-term psychotherapy. Manuscript submitted for publication.

Nisbett, R., & Ross, L. (1980). Human inference." Strategies and shortcomings of social judgment. Englewood Cliffs, NJ: Prentice-Hall.

Orlinsky, D., Grawe, K, & Parks, B. (1994), Process and outcome in psychotherapy - noch einmal. In A. Bergin & S. Garfield (Eds.), Handbook of psychotherapy and behavior change (4th ed.), (pp. 270-376). New York: Wiley.

Patterson, C. H. (1984). Empathy, warmth, and genuineness in psychotherapy: A review of reviews. Psychotherapy, 21, 431-438.

Rogers, C. R. (1957). The necessary and sufficient conditions of therapeutic personality change. Journal of Consulting and Clinical Psychology, 21, 95-103.

Safran, J. (1990). Towards a refinement of cognitive therapy in light of interpersonal theory: I. Theory. Clinical Psychology Review, 10, 87-105.

Safran, J., McMain, S., Crocker P., & Murray, E (1990). Therapeutic alliance rupture as a therapy event for empirical investigation. Psychotherapy, 27, 154-165.

Safran, J., & Segal, Z. (1990). Interpersonal process in cognitive therapy. New York: Basic. Safran, J., & Wallner, L. (1991). The relative predictive validity of two therapeutic alliance measures

in cognitive therapy. Psychological Assessment." A Journal of Consulting and Clinical Psychology, 3, 188-195. Scott, J., Byers, S., & Turkington, D. (1993). The chronic patient. In J. H. Wright, M. E. Thase,

A. T. Beck, &J. Ludgate (Eds.), Cognitive therapy with inpatients: Developing a cognitive milieu (pp. 357-390). New York: Guilford.

Spencer, S., & Hemmer, R. (1993). Therapeutic bias with gay and lesbian clients: A functional analysis, the Behavior Therapist 16, 93-97.

Stiles, W. B., Shapiro, D. A., & Elliot, R. (1986). Are all psychotherapies equivalent? American Psy- chologist, 42, 165'i80.

Strupp, H. 1988. What is therapeutic change?Journal of Cognitive Psychotherapy." An International Quar- terly, 2, 75-82.

Suh, C , O'Malley, S. S., Strupp, H., &Johnson, M. (1989). The Vanderbilt Psychotherapy Process Scale (VPPS). Journal of Cognitive Psychotherapy: An International Quarterly, 3, 123-154.

Suh, C., Strupp, H., & O'Malley, S. (1986). The Vanderbilt process measures: The Psychotherapy Process Scale (VPPS) and the Negative Indicators Scale (VNIS), In L. Greenberg & W. Pinsof (Eds.), The psychotherapeutic process: A research handbook (pp. 285-323). New York: Guilford.

Svartberg, M., & Stiles, T. (1992). Predicting patient change from therapist competence and patient- therapist complementarity in short-term anxiety-provoking psychotherapy: A pilot study. Journal of Consulting and Clinical Psychology, 60, 304-307.

Thase, M. E., & Wright, J. H. (1991). Cognitive behavior therapy manual for depressed inpatients: A treatment protocol oudine. Behavior Therapy, 22, 579-595.

Turkat, L D, & Maisto, S. A. (1985). Personality disorders: Application of the experimental method

THERAPEUTIC RELATIONSHIP 45

to the formulation and modification of personality disorders. In D. Barlow (Ed.), Clinical hand- book of psychological disorders." A step by step treatment manual (pp. 520-570). New York: Guilford Press.

Vessey, J. T., & Howard, K. I. (1993). Who seeks psychotherapy? Psychotherapy, 30(4), 546-553. Williams, K. E., & Chambless, D. (1990). The relationship between therapist characteristics and

outcome of in vivo exposure treatment for agoraphobia. Behavior Therapy, 21, 111-116. Wright, J. H. (1987). Cognitive therapy and medication as combined treatment. In A. Freeman

& V. Greenwood (Eds.), Cognitive therapy: Application in psychiatric and medical settings (pp. 36-50). New York: Human Sciences Press, Inc.

Wright, J. H. (1988). Cognitive therapy of depression. In A. J. Frances & R. E. Hales (Eds.), The American Psychiatric Press Review of Psychiatry (Vol 7, pp. 554-570). American Psychiatric Press, Inc: Washington, D.C.

Wright, J. H., & Beck, A. T. (in press). Cognitive therapy. In R. E, Hales, S. C. Yudofsky, & J. Talbott (Eds.), American Psychiatric Press 7kxtbook of Psychiatry. American Psychiatric Press: Wash- ington, D.C.

Wright, J. H., & Schrodt, G. R. (1989). Combined cognitive therapy and pharmacotherapy. In A. Freeman, M. K. Simon, H. Arkowitz, & L. Beutler (Eds.), Handbook of cognitive therapy (pp. 267-282). New York City: Plenum Press.

The authors would like to thank the Alliant Health System for sponsoring a keynote address by the second author at their 1993 Mental Health Conference. A portion of this address was used in the development of this paper. We also want to thank the three anonymous reviewers. Correspon- dence should be addressed to Denise Davis, Mood Disorders Clinic, The Village at Vanderbilt, 1500 21st Avenue South, Suite 2200, Nashville, TN 37212.

RECEIVED: September 20, 1993 ACCEPTED: March 24, 1994